[MUD-Dev] Classless systems

Steve Houchard stevehou at engin.umich.edu
Sun Jul 23 03:27:20 New Zealand Standard Time 2000

On Sat, 22 Jul 2000, David Bennett wrote:

> On Sat, 22 Jul 2000, Steve Houchard wrote:
> > If you only allow one variable to change then you're stuck with a one
> > dimensional character progression.  I'm personally quite fond of multiple
> > dimensions of progression.  Tough to balance?  Sure.  I think it's
> > worthwhile though because it opens up many more possibilities.  
> I don't agree.  

No one is asking you to agree :)

> Skills are significantly varied already and allow for
> much greator depth than any stat adjustment is likely to achive.  Unless
> you skill tree is amazingly small.  

If you have the same number of skills over both, how can limiting one side
allow for greater depth?  Even if the side that can vary stats has half
the skills as the other, it's still unclear to me how best to judge which
side has greater depth.  

> I would say that adjust stats would
> tend to make characters even more the same, since they can then start
> to adjust all of their stats to be the best possible person ever.

I probably didn't explain it well enough, but the cost of training
attributes starts prohibitively high and only becomes feasible when skills
reach the point of being prohibitively costly to train.  No low to
mid-level character will have the experience resources to train many of
their attributes more than one level or so.  Even very high level
characters would be hard pressed to regularly train an attribute.  At some
point it becomes far too expensive.  

> It is also highly unrealistic :)  

I fail to see the reasoning behind this statement.  

> Some things can be trained to be
> better over time, some things cannot.  I think you can simulate all
> the possible ways you can train things to be better by using skills
> rather than stats.  If you train to become better at beating people
> up with weapons then your overall stamina for doing this sort of
> thing will increase.  If you train yourserlf in learning languages it
> is as if y ou effectively became more inteligent.  Etc.

I studied Latin in college for many years before switching to Computer
Engineering.  In no way do I doubt that my success in Latin contributed to
my success in technical fields.  This can be modelled a few different
ways.  My choice is training the attribute which increases the overall
skill level in both the individual skills on top of training the
individual skills.  Other ways to accomplish this are skill webs and
other forms of intra-skill dependency.  Regardless of how you implement it
or explain it, you have the same thing; progression over multiple

I will agree that you can get the same range of numbers with a purely
skill-based progression scheme as with a multi-dimensional progressions
scheme.  I must disagree, however, that you can get the same level of
expressivity in much the same way as you might want to roll 2d6 rather
than 1d12.  

> > What I've done is implemented a monotonically increasing progression cost;
> > each successive level costs more than the last.  As a result, advancement
> > can get expensive rather quickly but it keeps the numbers within a
> > manageable frame.  Skills and stats are then on a sliding scale somewhere
> > on the cost axis with stats and some skills being more expensive than
> > others.  Furthermore, stats start on a base 10 scale typically around 6 or
> > 7 while skills will tend to start much lower.  
> > The end result is a theoretically unbounded progression scheme along many
> > dimensions that's in practice bounded by time (sometime after level 15 or
> > so, it becomes way to difficult to advance along the character's primary
> > skill set).  
> All of which you can achieve just using skills as well, with just
> as much variation (possibly more variation since people probably
> choose different start stats and cannot change them to be the
> same).  

How many ways can you express the number 11 on one scale?  How many ways
can you express it on 4 scales?  That, in a nutshell, is my argument.  If
the goal is variation in infinity limit values I can certainly see where
you're coming from.  Otherwise, I'm at a loss.  


MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list